Yannic Kilcher Courses

Regularizing Trajectory Optimization with Denoising Autoencoders - Paper Explained

Explore regularizing trajectory optimization in reinforcement learning using denoising autoencoders to improve planning and sample efficiency in motor control tasks.

• YouTube
• 30 minutes
• On-Demand
• Free Video

Automated Inference on Criminality Using Face Images

Critical analysis of a controversial facial recognition study, exploring ethical concerns and methodological flaws in attempts to predict criminality from facial features.

• YouTube
• 33 minutes
• On-Demand
• Free Video

Faster Neural Network Training with Data Echoing - Paper Explained

Explore data echoing technique to optimize neural network training by reusing pipeline data, reducing bottlenecks and improving GPU efficiency for faster processing.

• YouTube
• 39 minutes
• On-Demand
• Free Video

Group Normalization - Paper Explained

Explore Group Normalization as an alternative to Batch Normalization, addressing batch size limitations and improving performance in various deep learning tasks.

• YouTube
• 29 minutes
• On-Demand
• Free Video

Concept Learning with Energy-Based Models

Explore energy-based models for concept learning, focusing on their ability to generalize, reason abstractly, and solve problems creatively with limited data through inference-time optimization.

• YouTube
• 39 minutes
• On-Demand
• Free Video

Big Transfer - General Visual Representation Learning

Explore pre-trained ResNet for visual tasks, learning about large model pre-training and effective fine-tuning techniques for improved performance across diverse datasets.

• YouTube
• 34 minutes
• On-Demand
• Free Video

Divide-and-Conquer Monte Carlo Tree Search for Goal-Directed Planning - Paper Explained

Explores a novel AI planning approach using divide-and-conquer Monte Carlo Tree Search, proposing intermediate goals to break complex problems into manageable sub-tasks for more efficient problem-solving.

• YouTube
• 26 minutes
• On-Demand
• Free Video

Reinforcement Learning with Augmented Data

Explore how data augmentation enhances reinforcement learning algorithms, improving efficiency and generalization across visual observation tasks without modifying core RL methods.

• YouTube
• 22 minutes
• On-Demand
• Free Video

TAPAS - Weakly Supervised Table Parsing via Pre-training

Explore TAPAS, an innovative approach to table parsing and question answering using pre-training and weak supervision, outperforming traditional semantic parsing models with a simpler architecture.

• YouTube
• 32 minutes
• On-Demand
• Free Video

Chip Placement with Deep Reinforcement Learning

Deep reinforcement learning optimizes chip placement, outperforming human experts in speed and efficiency. This AI-driven approach learns from experience, generalizes to new designs, and minimizes power, performance, and area constraints.

• YouTube
• 27 minutes
• On-Demand
• Free Video

Jukebox - A Generative Model for Music

Explore OpenAI's Jukebox, a groundbreaking AI model generating high-fidelity music with vocals, controllable by genre, artist, and lyrics. Learn about its architecture and capabilities.

• YouTube
• 34 minutes
• On-Demand
• Free Video

The AI Economist - Improving Equality and Productivity with AI-Driven Tax Policies

Explore AI-driven tax policies that balance equality and productivity through reinforcement learning, revealing emergent strategies and outperforming traditional economic models in simulations and human experiments.

• YouTube
• 35 minutes
• On-Demand
• Free Video

Deconstructing Lottery Tickets - Zeros, Signs, and the Supermask

Explores the Lottery Ticket Hypothesis, analyzing key components of sparse networks and uncovering insights on weight initialization, sign importance, and the concept of Supermasks in neural network training.

• YouTube
• 36 minutes
• On-Demand
• Free Video

Do ImageNet Classifiers Generalize to ImageNet? - Paper Explained

Exploring ImageNet classifiers' generalization capabilities through new test datasets, revealing surprising accuracy drops and insights into model performance on slightly "harder" images.

• YouTube
• 26 minutes
• On-Demand
• Free Video

Supervised Contrastive Learning

Explore a novel supervised learning approach that outperforms cross-entropy loss, improving image classification models' performance and robustness across various architectures and data augmentations.

• YouTube
• 30 minutes
• On-Demand
• Free Video